Search and filtering should help users find content that meets their needs. That takes more than adding a search box or category menu. Search results should match the query, and titles and summaries should help users assess them. Filters should suit the content type and support the selection options users actually need. Both depend on accurate content tagging, and the scope of the search must be clear. Simply adding these features does not guarantee better usability.
Make results relevant, then make them understandable
A 2016 NN/g article noted that users may prefer searching, already know what they are looking for, or turn to search after other methods fail. In every case, results should match the query rather than leave users guessing.
- Titles and summaries should provide useful information to help users determine whether the content meets their needs.
- If search covers only part of a website, that limitation should be clearly disclosed so users do not mistakenly think they have searched the entire site.
- Ads mixed into search results may cause users to leave the site unintentionally. Designers should not overlook how this disrupts the search process.
Relevance also depends on content tagging. The article reported that, for a chicken recipe query on the AARP website, six of the first ten results were recipes without chicken, confusing users. Even a comprehensive filtering interface cannot replace accurate, reliable tagging.
Adapt filters to the content and task
The same set of filters will not suit every type of content. Finding a presentation about sustainability from last Tuesday may require narrowing results by file type, date, and topic at the same time. Someone buying shoes may care about heel height, while someone buying a coat may care about warmth. Filters should reflect the differences that matter to users, rather than offer the same criteria on every page.
Filters should also support different selection needs. A furniture shopper may want to see only blue chairs, or may want to exclude blue ones. The feature should support both needs.
Do not split a complete request across multiple queries
In the Maplin example documented by NN/g at the time, a user looking for Bluetooth speakers under £50 had to check separate, narrow price bands one by one and remember the products shown each time. The problem was not the absence of a price filter, but its inability to display all results meeting the full criteria at once. When designing filters, check whether users can directly specify the range they need without having to piece together the answer themselves.
Evaluate features using query logs and feedback
Regularly reviewing search logs can reveal how users search and which queries return suitable results. These findings can guide adjustments to content tagging and search features, making it easier for users to find results that match their queries.
A 2018 article on neilpatel.com provided another example involving a blog: after receiving feedback that content was hard to find, the author added topic filters and used heatmaps and scroll maps to examine how the filters were used and where they were placed. This experience on a specific page can inform evaluation methods, but it does not guarantee the same results on other websites.
How search and filtering should support users: perspectives from across the web
What signals can help assess usage after content filters are added?
User feedback, the distribution of attention across the page, and bounce rate can be examined together, but no single metric should determine whether the feature is effective. In a blog example shared by an author on neilpatel.com in 2018, the page's bounce rate fell by 21% after topic filters were added, and the filtering feature received 70% of the attention on the page. These observations apply to that page; they do not prove that filtering alone caused the decline or guarantee similar performance on other websites.
Can difficulties with search or filtering affect whether users stay on an e-commerce website?
They may. Data cited in a 2024 article on neilpatel.com claimed that four in five US shoppers had left an e-commerce website because its on-site search or filtering was frustrating. This claim covers search or filtering and does not establish how many departures each caused separately. The supplied material also did not explain the survey methodology, so the proportion should not be generalized to all websites or users.
References
- How I Boosted My Rankings Without Creating Content or Building Linksneilpatel.com · 2018-08-28T13:52:53+00:00 — Discusses: What signals can help assess usage after content filters are added?
- Top 10 Enduring Web-Design Mistakeswww.nngroup.com · 2016-10-30T16:00:02+0000 — Source used in the article above.
- The Power of Search Experience Optimization (SXO)neilpatel.com · 2024-04-29T19:00:00+00:00 — Discusses: Can difficulties with search or filtering affect whether users stay on an e-commerce website?


